I would like to understand how the standard error for the median survival time is calculated in SPSS 19.0
I've looked at the Algorithms document (ftp://ftp.software.ibm.com/software/analytics/spss/support/Stats/Docs/19.0/Client/User_Manuals/English/IBM_SPSS_Statistics_19_Algorithms.pdf) and couldn't find an answer.
I found this review article (http://www.barkerstats.com/PDFs/tas.pdf) which outlines the formula in SPSS 17.0
$$ se(P_i) = P_i \sqrt{\sum^i_{j=1} ( \frac{q_j}{r_j p_j})} $$
Is this correct? If yes, how do I properly apply this quantity to the median survival time?
